Jökull - 01.12.1989, Síða 37

in Reykjavík as a function of time. The instrument
was sent for recalibration and adjustment after the
1982 measurement. Vertical bars show standard
deviation determined from several observations
within hours of each other.
Mynd 3. Mœld lengd á viðmiðunarlínu 1 í Reykjavík
sýnd sem fall af tíma. Talsvert frávik kemur fram í
mœlingunni 1982 og var tœkið þá sent til framleið-
anda til stillingar. Hver punktur táknar meðaltal
nokkurra mœlinga sem gerðar eru innan fárra
klukkutíma. Lóðrétt strik sýna staðalfrávik.
this correction were determined from the atmos-
pheric pressure differences.
A 1454 m long line in a presumably stable area in
Reykjavík was measured several times during the
experiment (Fig. 3) to monitor the calibration of the
instrument and allow correction for systematic
errors such as changes in the instrument constant
and modulation frequency. The variation in the
measured length is small except for the 1982 value.
The instrument was sent for repair and recalibration
after this measurement. A second, longer calibra-
tion line was installed in 1985 to get an idea of the
repeatability of the results under various meteoro-
logical conditions (Fig.4). Several measurements
were done during a few weeks interval and in
different weather. The total range of values was
32 mm, and the standard deviation was 7 mm.
M
JUN
1985
CALIBRATION LINE 2
Fig. 4. The lengths of the calibration lines measured
under various meteorological conditions in May
and June 1985. Each point represents several read-
ings within hours of each other. Vertical bars show
the standard deviation of the observations made.
Mynd 4. Lengd viðmiðunarlína mœld í mismunandi
veðri í maí og júní 1985. Hver punktur táknar með-
altal nokkurra mœlinga innan fárra klukkutíma.
Lóðrétt strik sýna staðalfrávik mœlinganna.
Corresponding values for the short calibration line
are 25 mm and 6 mm. The total range of the aver-
ages shown in Figs. 3 and 4 was 15 mm and 14 mm
JÖKULL, No. 39, 1989 35